• Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法

​
​よろず知財コンサルティングのブログ

AIによる支援は専門知識を強化するのか、損なうのか?

8/11/2026

0 Comments

 
「AIによる支援は専門知識を強化するのか、それとも損なうのか?特許明細書作成における3ヶ月間の実地実験からの証拠」というワーキングペーパーは、マサチューセッツ工科大学のデビッド・オーター教授やグーグルなどの研究チームが実施した、特許弁護士133人を対象とする3ヶ月間のランダム化比較試験に関する実証研究の報告です。この研究は、生成AIツール「InFlow」の業務導入が、AI支援下における目先の業務パフォーマンスと、AIを取り去った後に個人の頭脳へ定着する長期的な専門能力の双方にどのような影響を与えるかを検証したものです。「AI支援下の能力」と「AI撤去後の独立した専門的判断力」を分離して測定したところに特徴があります。
実験の結果、AIの利用により若手弁護士は作業時間の短縮や品質の底上げという大きな恩恵を受けた一方で、AIを禁止した状態での査読課題ではベテラン層のみが能力の向上を示し、若手層の間ではスキルの二極化が生じるという非対称性が確認されました。
この研究について生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
DOES AI ASSISTANCE ENHANCE OR ERODE EXPERTISE? EVIDENCE FROM
A THREE-MONTH FIELD EXPERIMENT IN PATENT DRAFTING
AIによる支援は専門知識を強化するのか、それとも損なうのか?特許明細書作成における3ヶ月間の実地実験からの証拠
https://www.nber.org/system/files/working_papers/w35720/w35720.pdf
2026年11月08日
AIはベテランの判断力を鍛え、若手の思考を代替しうる — 特許弁護士133人・3ヶ月のランダム化比較試験が示した分岐点
https://techfeed.io/entries/6aefaa6649a7d098b464cd7b
Prior expertise may determine whether AI builds skills or substitutes for them, field experiment suggests
https://www.psypost.org/prior-expertise-may-determine-whether-ai-builds-skills-or-substitutes-for-them-field-experiment-suggests/
 
 
Does AI Assistance Strengthen or Erode Expertise?
The working paper “Does AI Assistance Strengthen or Erode Expertise? Evidence from a Three-Month Field Experiment in Patent Drafting” presents empirical findings from a three-month randomized controlled trial involving 133 patent attorneys, conducted by a research team that included Professor David Autor of the Massachusetts Institute of Technology and researchers from Google, among others. The study examined how introducing the generative AI tool “InFlow” into professional workflows affected both immediate job performance with AI assistance and the lasting professional skills individuals retained after AI support was withdrawn. Its distinctive feature lies in measuring “performance with AI assistance” separately from “independent professional judgment after AI support was withdrawn.”
The experiment revealed an asymmetry: junior attorneys benefited substantially from AI through reduced task completion times and improvements in work quality, but only experienced attorneys demonstrated improvements in their abilities on peer-review tasks in which AI use was prohibited. Meanwhile, skill outcomes became polarized among junior attorneys.
I asked generative AI to conduct an in-depth analysis of this study and invite you to read the results. Please bear in mind that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual circumstances, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

旭化成がAI活用の「スマートラボラトリー」を開設

7/11/2026

0 Comments

 
2026年11月5日、旭化成は、静岡県富士市に、AI(人工知能)や自動化技術を活用し、材料開発の一連のプロセスを自律的に行うプラットフォーム「スマートラボラトリー」を開設すると発表しました。2027年1月から稼働予定で、半導体向け感光性絶縁材料「パイメル」の製品開発を加速させるということです。
この発表を生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
旭化成がAI活用のスマートラボ新設、半導体材料の実験数4倍に
https://news.yahoo.co.jp/articles/251e9c2eb6eaee9cf4660fd16e558cba2acb79f8
 
 
Asahi Kasei to Open an AI-Powered “Smart Laboratory”
On November 5, 2026, Asahi Kasei announced that it would open a “Smart Laboratory” in Fuji City, Shizuoka Prefecture. The platform will use artificial intelligence (AI) and automation technologies to autonomously carry out a series of materials development processes. Scheduled to begin operations in January 2027, it is intended to accelerate product development for PIMEL, the company’s photosensitive insulating materials for semiconductor applications.
I asked generative AI to take an in-depth look at this announcement. Please refer to the resulting research and analysis, bearing in mind that these findings are based solely on publicly available information, do not necessarily reflect actual circumstances, and may contain errors.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

書籍『他社特許の調査・分析法と障害特許の崩し方、攻め方』

6/11/2026

0 Comments

 
私も一部を執筆した書籍『他社特許の調査・分析法と障害特許の崩し方、攻め方』が、2026年10月30日に技術情報協会から発刊されました。
本書は、他社特許の調査・分析から、生成AIの活用、特許権侵害の回避、無効化、判例を踏まえた権利解釈、ライセンス交渉、侵害警告への対応まで、他社の知的財産に向き合う実務を幅広く扱っています。特許に加え、商標・意匠の調査や侵害回避も取り上げられています。
私が担当したのは、第1章第6節「参入障壁となる先発企業の特許網の特定と弱みの見つけ方――特許ランドスケープ,クレーム分析,AI活用による実践的アプローチ」です。
新たな市場への参入や新製品の開発では、先発企業が築いた特許網への対応が重要な課題になります。そこで必要になるのが、特許網の全体像を把握し、自社の製品・技術との関係を具体的に検討することです。
担当節では、IPランドスケープによる可視化、引用ネットワークや審査経過の分析、クレームチャートによる請求項と自社製品の対比などを通じて、重要な特許を見極める進め方を整理しました。そのうえで、特許網の弱みを探る視点を、法的有効性、権利範囲、時系列、国別・市場別、維持状況・事業組織上の事情の五つに分けて解説しています。
例えば、権利範囲に含まれない技術構成、主要特許の満了時期、国ごとの権利取得状況などを確認することは、回避設計や参入時期・市場の選択を検討する手がかりになります。こうした分析を、無効化、ライセンス交渉、自社特許網の構築といった具体的な対応につなげる考え方を取り上げました。
生成AIやAIエージェントの活用についても、特許情報の分類・要約、請求項の整理、検索語の展開、回避案の検討など、分析の各工程に沿って紹介しています。
あわせて、AIの出力を原文や出典と照合すること、機密情報の扱いを確認すること、侵害・有効性・事業リスクに関する判断を人が担うことなど、実務に組み込む際の役割分担も整理しました。知財・研究開発・事業部門が連携し、調査・分析の成果を具体的な事業判断につなげていくための視点です。
他社特許への対応に携わる知財部門や弁理士の方々、研究開発や新規事業を担当される方々に、日々の課題を検討する際の参考としてご活用いただければ幸いです。
 
書籍情報
書名:他社特許の調査・分析法と障害特許の崩し方、攻め方
発行:技術情報協会
発刊日:2026年10月30日
体裁:A4判・574頁
定価:製本版・USB版ともに88,000円(税込)
 
書籍の詳細な目次や試読の案内は、技術情報協会の書籍紹介ページをご覧ください。
 
他社特許の調査・分析法と障害特許の崩し方、攻め方
https://www.gijutu.co.jp/doc/b_2378.htm
 
Searching and Analyzing Third-Party Patents: How to Challenge and Overcome Blocking Patents
The book Searching and Analyzing Third-Party Patents: How to Challenge and Overcome Blocking Patents, to which I contributed, was published by Technical Information Institute Co., Ltd. on October 30, 2026.
The book covers a broad range of practical issues involved in dealing with third-party intellectual property, including patent searching and analysis, the use of generative AI, infringement avoidance, patent invalidation, interpretation of patent scope in light of case law, licensing negotiations, and responses to infringement warning letters. In addition to patents, it addresses trademark and design searches and infringement avoidance.
I contributed Chapter 1, Section 6, “Identifying Incumbents’ Patent Portfolios That Create Barriers to Entry and Finding Their Weaknesses—A Practical Approach Using Patent Landscaping, Claim Analysis, and AI.”
When entering a new market or developing a new product, addressing the patent portfolios built by established companies is a major challenge. This requires an overall understanding of those portfolios and a detailed assessment of how they relate to one’s own products and technologies.
In my section, I outline an approach to identifying key patents through visualization using IP landscape analysis, examination of citation networks and prosecution histories, and claim charts that compare patent claims with a company’s own products. I then explain five perspectives for identifying weaknesses in patent portfolios: legal validity; scope of protection; changes over time; differences by country and market; and patent maintenance status and business or organizational circumstances.
For example, examining technical configurations that fall outside the scope of protection, the expiration dates of key patents, and the status of patent protection in each country can help inform design-around strategies and decisions about when and where to enter a market. I discuss how such analysis can lead to concrete actions, including pursuing patent invalidation, negotiating licenses, and building one’s own patent portfolio.
I also introduce ways to use generative AI and AI agents at each stage of the analysis, including classifying and summarizing patent information, organizing claim content, expanding search terms, and exploring design-around options.
In addition, I outline the division of responsibilities when integrating AI into practice. This includes checking AI outputs against original texts and sources, verifying how confidential information is handled, and ensuring that people remain responsible for judgments concerning infringement, patent validity, and business risks. These considerations support collaboration among IP, R&D, and business departments, helping translate search and analysis findings into concrete business decisions.
I hope this book will serve as a useful reference for IP departments and patent attorneys handling third-party patent issues, as well as those involved in R&D and new business development, as they work through the challenges they face in their daily practice.
Book Details
  • Title: Searching and Analyzing Third-Party Patents: How to Challenge and Overcome Blocking Patents
  • Publisher: Technical Information Institute Co., Ltd.
  • Publication date: October 30, 2026
  • Format: A4, 574 pages
  • List price: JPY 88,000 each for the print and USB editions, including tax
For the detailed table of contents and information on preview options, please visit the publisher’s book information page.

Picture
0 Comments

攻めの知財へ「特許の怪物」も使いよう?

6/11/2026

0 Comments

 
『攻めの知財へ「特許の怪物」も使いよう 訴訟恐れぬ日本企業に』という日経新聞の署名入り記事を読み、驚きました。
生成AIに、肯定論、否定論を探らせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
攻めの知財へ「特許の怪物」も使いよう 訴訟恐れぬ日本企業に
https://www.nikkei.com/article/DGXZQOCD027UW0S6A001C2000000/
 
 
Taking the Offensive in IP: Can Even “Patent Trolls” Be Put to Good Use?
I was surprised to read a bylined article in Nikkei titled “Taking the Offensive in IP: Even ‘Patent Trolls’ Can Be Put to Good Use—A Call for Japanese Companies Not to Fear Litigation.”
I asked generative AI to explore arguments both for and against this approach. The findings are presented for your reference.
Please note that the research and analysis generated by AI are based solely on publicly available information, do not necessarily reflect the actual state of affairs, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

パラメータ発明の拒絶理由への対応策

5/11/2026

0 Comments

 
月刊パテント9月号掲載の「中国審査実務におけるパラメータ発明に関わる拒絶理由への対応策」(『パテント』2026年、Vol.79、No.9、51–62頁、朱丹・王銘浩)の論点について、生成AIに、日本の審査実務と比較させました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
中国審査実務におけるパラメータ発明に関わる拒絶理由への対応策
https://link.directcloud.jp/4pgkmsqh5G
 
 
Strategies for Responding to Grounds for Rejection of Parameter Inventions
I asked generative AI to compare Chinese and Japanese patent examination practices regarding the issues discussed in “Strategies for Responding to Grounds for Rejection of Parameter Inventions in Chinese Patent Examination Practice,” by Zhu Dan and Wang Minghao, published in the September issue of the monthly journal Patent (2026, Vol. 79, No. 9, pp. 51–62).
When reviewing the results, please bear in mind that the research and analysis conducted by generative AI are based solely on publicly available information, do not necessarily reflect actual circumstances, and may contain errors.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

令和8年〔ネ〕第10023号 意識的除外

4/11/2026

0 Comments

 
知財高裁令和8年8月27日判決(令和8年〔ネ〕第10023号)は、電気コーヒーメーカーの特許侵害を否定した原判決を維持し、特許権者の控訴を棄却しました。注目したいのは、権利化のための補正や意見書の説明が、後の権利行使で不利に働いた点です。
特許権者は、本件特許2の拒絶理由への対応で、ヒータを「中空に吊り下げ」る構成を追加し、意見書で「ボイラー底全面が給湯孔のために使用できる」と説明しました。裁判所は、この出願経過に照らし、底面に接触して固定された被告製品のヒータ構成は意識的に除外されたとして、均等侵害を否定しました。
本件の教訓は、単に意見書の説明を減らすことではなく、補正と説明によって、どの代替構成を権利範囲から除外したと理解されるかを点検することです。「取れる特許」と「使える特許」を一体で考える必要性を示す事例です。
この判決について、判決文や既存の評釈を基に、複数の生成AIに実務上の示唆を検討させましたので、ご参照ください。なお、生成AIによる分析には誤りが含まれる可能性があります。判決の認定・判断と評釈の評価を区別し、原文も確認のうえ、ご参照ください。
 
令和8年(ネ)第10023号 特許権侵害差止等請求控訴事件 判決
https://www.courts.go.jp/assets/hanrei/hanrei-pdf-96986.pdf
 
Case No. 2026 (Ne) 10023: Intentional Exclusion
In its judgment of August 27, 2026 (Case No. 2026 (Ne) 10023), the Intellectual Property High Court upheld the lower court’s judgment of non-infringement in a case concerning an electric coffee maker and dismissed the patentee’s appeal. Of particular interest is how amendments and explanations submitted during prosecution to secure patent protection later worked against the patentee when enforcing its rights.
In responding to the reasons for refusal concerning Patent 2 at issue, the patentee added a limitation requiring the heater to be “suspended in midair” and explained in its written submissions that “the entire bottom surface of the boiler can be used for hot-water supply openings.” In light of this prosecution history, the court held that the heater configuration in the defendant’s products—in which the heater was fixed in contact with the boiler’s bottom surface—had been intentionally excluded from the scope of protection, and therefore found no infringement under the doctrine of equivalents.
The lesson of this case is not simply to provide fewer explanations in written submissions, but to examine which alternative configurations may be understood to have been excluded from the scope of protection by the amendments and accompanying explanations. The case illustrates the need to consider “patents that can be obtained” and “patents that can be effectively enforced” together.
Drawing on the judgment and existing case commentaries, I asked several generative AI systems to examine the practical implications of this decision. Please refer to their analyses, bearing in mind that analyses generated by AI may contain errors. When reviewing them, please distinguish the court’s findings and rulings from the assessments offered in the commentaries, and consult the original judgment as well.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

カバンに入る「AIスパコン」を手元で動かす

4/11/2026

0 Comments

 
YouTube「【カバンに入る「AIスパコン」を手元で動かす】NVIDIAがWindows専用チップを開発→GPT級AIがPCで使い放題/クラウド不要でサイバー防衛/HP 世界の事業責任者に直撃【1on1 Tech】」は、HPパーソナルシステムズの責任者であるカイトン・パテル氏へのインタビューを通じ、人工知能の進化がパーソナルコンピュータの役割を根本から変革している現状が詳しく解説されています。
従来のPCは人間の指示に対して結果を返すだけの受動的なツールでしたが、今後はローカル環境で自律的に稼働する多数のAIエージェントと共存するパーソナルインテリジェンス機器へと進化します。その結果、クラウド依存に伴う高額なトークン費用や機密データの漏洩リスクといった課題がクリアされます。
生成AIに、【カバンに入る「AIスパコン」を手元で動かす】に関して深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
YouTube「【カバンに入る「AIスパコン」を手元で動かす】NVIDIAがWindows専用チップを開発→GPT級AIがPCで使い放題/クラウド不要でサイバー防衛/HP 世界の事業責任者に直撃【1on1 Tech】」
https://www.youtube.com/watch?v=DT-G-T3gkEk
 
 
Running an “AI Supercomputer” That Fits in Your Bag
The YouTube video “[Running an ‘AI Supercomputer’ That Fits in Your Bag] NVIDIA Develops a Windows-Specific Chip → Unlimited Use of GPT-Class AI on Your PC / Cyber Defense Without the Cloud / An Exclusive Interview with HP’s Global Business Leader [1on1 Tech]” provides an in-depth look at how advances in artificial intelligence are fundamentally transforming the role of personal computers, through an interview with Ketan Patel, head of HP’s Personal Systems business.
Traditional PCs have been passive tools that simply return results in response to human instructions. In the future, however, they will evolve into personal intelligence devices that host numerous AI agents operating autonomously and locally. This evolution will resolve challenges associated with reliance on the cloud, such as high token costs and the risk of confidential data leaks.
I asked generative AI to explore the topic of “Running an ‘AI Supercomputer’ That Fits in Your Bag” in greater depth. Please refer to the resulting research and analysis, bearing in mind that the findings are based solely on publicly available information, may not necessarily reflect the actual situation, and could contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

AIが科学研究を加速させている

3/11/2026

0 Comments

 
科学論文のプレプリントサーバー「arXiv」で、数学論文の投稿数が1年で約2倍に跳ね上がる劇的な変化が起きています。YouTube「【数学論文が1年で2倍】AIが科学研究を加速させている」によれば、背景にあるのは、データ収集から分析、執筆まで自律して進めるAIエージェントの存在だということです。
この「AIが科学研究を加速させている」状況を生成AIに深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
【数学論文が1年で2倍】AIが科学研究を加速させている
https://www.youtube.com/watch?v=CmH_iwRN3jo
 
 
AI Is Accelerating Scientific Research
A dramatic shift is taking place on arXiv, the preprint server for scientific papers, where the number of mathematics papers submitted has roughly doubled in just one year. According to the YouTube video “[Mathematics Papers Double in One Year] AI Is Accelerating Scientific Research,” this surge is being driven by AI agents that work autonomously through the research process, from data collection and analysis to writing.
I asked generative AI to take an in-depth look at how AI is accelerating scientific research, and I invite you to review its findings. Please note, however, that the research and analysis produced by generative AI are based solely on publicly available information, do not necessarily reflect the actual situation, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

ソフトバンクグループの特許出願・権利化動向

2/11/2026

0 Comments

 
ソフトバンクグループ株式会社の特許活動は、2023年秋以降、社内アイデアコンテストと生成AIを活用した前例のない「超大量出願」と、その後の「徹底した権利化選別(多産少死)」および「トップ主導案件への外国出願集中」という極めて特異な二極化構造を示しています。
このソフトバンクグループの特許出願・権利化動向を生成AIに深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
SoftBank Group’s Patent Filing and Prosecution Trends
Since the fall of 2023, SoftBank Group Corp.’s patent activities have exhibited a highly distinctive, polarized pattern: unprecedented “ultra-high-volume patent filing” driven by internal idea competitions and generative AI, followed by “rigorous selection of applications to pursue through to grant (‘high output, low attrition’)” and “a concentration of foreign filings on initiatives led by top management.”
I asked generative AI to conduct an in-depth analysis of these trends in SoftBank Group’s patent filing and prosecution activities. Please refer to the resulting report.
Please note, however, that the AI-generated research and analysis are based solely on publicly available information, do not necessarily reflect the actual circumstances, and may contain errors. Please bear these limitations in mind when reviewing the report.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

ブログ記事一覧(2026年10月)

2/11/2026

0 Comments

 
10月31日 特許庁業務変革ロードマップ(暫定版)
10月31日 中国の特許出願の現状をどうみるか
10月30日 「Gemini 4 Carbon」が知財業務へ与える影響
10月30日 国産生成AI基盤モデルPLaMo Prime
10月30日 中国特許におけるAI生成コンテンツ規制
10月29日 OpenAI「学術界と連携して科学的知見を構築する」
10月28日 OpenAI、「GPT-6.1 Terra」発表
10月28日 「HP ZGX Fury G1n AI Station」の知財業務への影響
10月27日 知財業務にAIエージェントを適用する際のハーネス設計
10月27日 AIエージェントのハーネス不要論
10月26日 国産生成AIと知財の最前線
10月26日 知財・無形資産ガバナンス協会の機関誌第2号
10月25日 Tesla、人型ロボット「Optimus」を量産ラインで生産
10月25日 「Kimi K3.1 Pro」が知能ベンチマーク世界4位
10月25日 「JIMI企業」に学ぶ無形資産を顧客価値と利益につなぐ経営
10月24日 花王、膨大な「研究データ」を生成AI で探索可能に。
10月23日 AIが自律的に創出した発明(AI自律発明)の保護
10月23日 AIが設計したウイルスをヒトに投与
10月22日 OpenAIがOrchestra Bioを買収
10月22日 最終判断は人間がやる
10月21日 データセンター・クラウド安全認証
10月21日 MiniMax M3.1 Flash の正式リリース
10月20日 ヒューマノイド・フィジカルAIで日本は勝てるのか
10月19日 大量の特許を継続的に扱う業務に強いPatentfield AIR
10月19日 ブログの紹介(2026年1月―2026年9月)
10月18日 調査・権利化の自動化に強いSummaria
10月18日 マルチモーダル軽量推論モデル「DeepSeek-V4.1-Flash-1017」
10月18日 トヨタSDV全面移行方針と知財への影響
10月17日 論文の矛盾を見つけるAI「Deep Review」
10月17日 AIで深化する特許実務の現場から
10月17日 「Muse Spark Open」により企業の知財業務がどう変わる
10月16日 IPランドスケープ・知財戦略立案に強いBiz Cruncher
10月15日 国産生成AI「Sarashina」を無料で試せるサービス
10月14日 「Gemini 4 Argon」による知財業務への影響
10月14日 「Gemini 4 Argon」が一般提供開始
10月14日 生成AIが変える業務と提供価値(旭化成)
10月14日 明確性要件の積極的適用――特許庁の近時実務に注目
10月14日 MetaがMuse Wonder and Muse Wonder Full発表
10月13日 Sakana AIによる「MASS」技術
10月13日 米国60億ドル科学イニシアチブと日米比較
10月12日 画像生成AIの新モデル「Nano Banana 2.1」の知財実務活用
10月12日 Google DeepMindの数学証明AIシステム「AlphaProof Nexus」
10月12日 再帰的自己改善を測るベンチマーク「Vals RSI Index」
10月12日 ドイツAI企業が発表したLLM「Kolibri」
10月11日 日本とフランスにおける政府AI基盤の比較
10月11日 令和6年(受)第2078号 損害賠償請求事件最高裁判決バンドスコア
10月11日 OpenAIの研究の80~90%は将来モデルが対象
10月11日 令和8年(行ケ)第10016号 手続違背
10月10日 NVIDIA RTX Sparkにおける知財用国産ローカルLLM
10月10日 NVIDIA RTX Sparkにおける知財特化型用ローカルLLM選定
10月10日 生成AIの知財業務における戦略的活用研修
10月10日 NVIDIA RTX Spark搭載PCが拓くLocal LLMと知財業務
10月 9日 Claude Haiku 5.5が企業の知財実務に与える影響
10月 9日 AnthoropicがClaude Haiku 5.5発表
10月 8日 OpenAIがAIによる722本の数学論文を公開
10月 8日 日本はAIで出遅れたが、それでも「国産AI」が必要な理由
10月 8日 Mistral Large 4が日本へ与える教訓
10月 8日 Mistral Large 4
10月 8日 自律型AIエージェントの大規模刷新版「Manus 2.0」
10月 7日 日本版AIセーフティ・インスティテュート(AISI)
10月 7日 商用特許SDIサービスの現状と今後
10月 7日 生成AIによる特許SDI調査の変化
10月 6日 高市首相所信表明演説における知財・無形資産投資
10月 6日 「知的財産が企業経営に与える影響に関する調査研究」
10月 5日 高市首相の所信表明演説
10月 5日 ロボットに「できる仕事」と「採算が合う仕事」は違う
10月 4日 WIPO Global Innovation Index 2026の日本企業知財業務への示唆
10月 4日 ELYZA-Thinking-1.0(llm-jp-4版)無料公開
10月 4日 過去のフロンティアモデルと現在のローカルLLMの比較
10月 3日 自律型AIエージェントの「脱走」・不正行動を防ぐ技術
10月 3日 日本成長戦略会議(第8回)が企業知財業務に与える影響
10月 3日 NVIDIAが「NVIDIA Open Agent Safety Platform」を発表
10月 2日 高市首相が高性能AI普及へ年内に行動計画策定指示
10月 2日 GoogleがGemini 4 Argonを発表
10月 1日 AI研究開発自動化に伴う知能爆発の知財業務への影響
10月 1日 AI研究開発自動化に伴う「知能爆発」
10月 1日 OpenAI DevDay 2026諸技術による知財業務への影響
 
 
過去のブログ
10月19日 ブログの紹介(2026年1月―2026年9月)
5月 8日 2025年7月1日~2025年12月31日 ブログ記事一覧
5月 8日 2025年1月1日~2025年6月30日 ブログ記事一覧
10月19日 ブログの紹介(2024年&2023年)
10月20日 ブログ紹介(2022年)
 
 

0 Comments

営業秘密を中国企業に漏洩した不正競争防止法違反

1/11/2026

0 Comments

 
名古屋地検は、2026年10月27日、旭化成の半導体材料に関する営業秘密を中国企業に漏洩したとして、不正競争防止法違反の罪で同社元社員を起訴しました。
この事件の内容を深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
情報漏えい「知識豊富と示すため」 元旭化成社員を起訴―名古屋地検
時事通信 社会部2026年10月27日19時40分配信
https://www.jiji.com/jc/article?k=2026102701077&g=soc#goog_rewarded
 
 
Alleged Trade Secret Leak to a Chinese Company in Violation of the Unfair Competition Prevention Act
On October 27, 2026, the Nagoya District Public Prosecutors Office indicted a former Asahi Kasei employee on charges of violating the Unfair Competition Prevention Act for allegedly disclosing the company’s trade secrets relating to semiconductor materials to a Chinese company.
I asked generative AI to examine this case in depth, and I invite you to review its findings. Please note, however, that the AI-generated research and analysis are based solely on publicly available information, do not necessarily reflect the actual circumstances, and may contain inaccuracies.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

特許庁業務変革ロードマップ(暫定版)

31/10/2026

0 Comments

 
令和8年10月30日、特許庁は、「特許庁業務変革ロードマップ(暫定版)」を公表しました。この「特許庁業務変革ロードマップ(暫定版)」について、生成AIに徹底的に調べさせ、米国、欧州、中国、韓国の特許庁と比較検討させましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
特許庁業務変革ロードマップ(暫定版)
https://www.jpo.go.jp/system/laws/sesaku/ai_action_plan/gyoumuhenkaku-roadmap/index.html
 
 
Japan Patent Office’s Operational Transformation Roadmap (Provisional Version)
On October 30, 2026, the Japan Patent Office (JPO) published its “Operational Transformation Roadmap (Provisional Version).” I asked generative AI to conduct an in-depth investigation of the roadmap and compare the JPO’s approach with those of the patent offices in the United States, Europe, China, and South Korea. The findings are provided for your reference.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, may not necessarily reflect the actual situation, and may contain inaccuracies. Please bear these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

中国の特許出願の現状をどうみるか

31/10/2026

0 Comments

 
中国の特許出願件数の急増については、「技術覇権を脅かす重大な脅威である」という議論と、まったく逆の「大半が国内向けの低質な特許(水増し)であり大したことはない」という議論があり、大きなギャップがあるようです。
生成AIに、中国の特許出願について、徹底的に比較・分析するように指示し、調査レポートを作成させましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
How Should We Assess the Current State of China’s Patent Filings?
There appears to be a wide divide in views on the rapid surge in China’s patent filings. Some argue that it poses “a serious threat to technological dominance,” while others take the opposite view, asserting that “most are low-quality patents aimed at the domestic market—merely inflating the numbers—and are not a major concern.”
I asked generative AI to conduct a thorough comparative analysis of China’s patent filings and prepare a research report, which is provided here for your reference.
Please bear in mind that the findings of this AI-generated research and analysis are based solely on publicly available information, do not necessarily reflect the actual situation, and may contain inaccuracies.
 

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

「Gemini 4 Carbon」が知財業務へ与える影響

30/10/2026

0 Comments

 
Googleは2026年10月28日(日本時間29日午前2時)、フロンティアモデル「Gemini 4 Carbon」を発表しました。サイバーセキュリティと長時間タスクの遂行を強みとしていて、Artificial Analysis Intelligence Index v5.0では、首位の59となっています。(2位はClaude Opus 5.5 で58、GPT-6 Astra は57)
今週から一部のサイバー防御担当者へ限定提供され、その後「Fairwind Program」を通じて信頼できるパートナーへ段階的に拡大されます。開発者や一般向けの提供時期は未定。
この「Gemini 4 Carbon」が知財業務へ与える影響について生成AIに調べさせました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5
https://artificialanalysis.ai/articles/gemini-4-carbon-new-number-1
 
 
The Impact of “Gemini 4 Carbon” on Intellectual Property Work
Google announced its frontier model, “Gemini 4 Carbon,” on October 28, 2026 (at 2:00 a.m. on October 29, Japan time). Its strengths include cybersecurity and the execution of long-running tasks. It ranks first on the Artificial Analysis Intelligence Index v5.0 with a score of 59. (Claude Opus 5.5 is in second place with 58, while GPT-6 Astra scores 57.)
Beginning this week, the model will be made available to a select group of cyber defense professionals, with access subsequently expanding in phases to trusted partners through the “Fairwind Program.” The timing of availability for developers and the general public has yet to be determined.
I asked generative AI to investigate the impact of “Gemini 4 Carbon” on intellectual property work.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

国産生成AI基盤モデルPLaMo Prime

30/10/2026

0 Comments

 
株式会社Preferred Networksは、国内でフルスクラッチ開発する生成AI基盤モデルPLaMo™の最新フラッグシップモデルPLaMo 3.1 Primeを、2026年10月27日より提供開始しました。新たに搭載したDeep Researchは、オンプレミス環境でも利用可能で、調査計画の立案から情報の収集・整理・分析、出典付きレポートの作成までを自律的に実行して、企業の調査業務を支援します。
国産生成AI基盤モデルPLaMo Primeについて、特に知財業務への利用について、生成AIに深堀させました。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
PLaMo 3.1 Primeをリリースしました
https://www.preferred.jp/ja/blog/tech/plamo-3-1-prime-release
 
国産生成AI基盤モデルPLaMo PrimeにWeb検索・Deep Research搭載 調査計画の立案から出典付きレポートまで自律的に実行、オンプレミスにも対応
https://www.preferred.jp/ja/news/pr20261027
 
 
PLaMo Prime: A Generative AI Foundation Model Developed in Japan
On October 27, 2026, Preferred Networks, Inc. began offering PLaMo 3.1 Prime, the latest flagship model in its PLaMo™ lineup of generative AI foundation models developed from scratch in Japan. The newly added Deep Research feature can also be used in on-premises environments. It supports companies’ research activities by autonomously carrying out the entire process, from developing research plans and collecting, organizing, and analyzing information to producing reports with source citations.
I asked generative AI to take an in-depth look at PLaMo Prime, a generative AI foundation model developed in Japan, with a particular focus on its applications in intellectual property work.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

中国特許におけるAI生成コンテンツ規制

30/10/2026

0 Comments

 
中国における特許出願の予備審査(専利預審)において、AIが生成した技術内容が含まれることが問題視され、各地の知的財産保護中心で審査の厳格化や罰則が科されているということです。
現状と今後について、生成AIに調べさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
Regulation of AI-Generated Content in Chinese Patent Applications
It has been reported that the inclusion of AI-generated technical content in patent applications is being flagged as problematic during China’s patent pre-examination (专利预审) process, with intellectual property protection centers in various regions tightening scrutiny and imposing penalties.
I asked generative AI to investigate the current situation and future outlook, and am sharing the findings for your reference. Please note that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual situation, and may contain errors. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

OpenAI「学術界と連携して科学的知見を構築する」

29/10/2026

0 Comments

 
2026年10月26日(米国現地時間)、OpenAIは、「Working with academic communities to build scientific understanding(学術界と連携して科学的知見を構築する)」と題する研究発表で、数学、計算機科学、物理学の各分野において、調和解析、計算量理論、場の量子論に関する 3 つの新たな発見を発表しました。このOpenAI「学術界と連携して科学的知見を構築する」発表の影響について生成AIに調べさせました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
2026年10月27日
OpenAIが数学・コンピュータサイエンス・物理学における調和解析・計算複雑性理論・量子場理論に関する3つの新たな発見を発表、指数時間仮説は誤りである
https://gigazine.net/news/20261027-openai-three-new-discoveries/
 
Working with academic communities to build scientific understanding | OpenAI
https://openai.com/index/building-scientific-understanding-with-academic-communities/
 
 
OpenAI: “Working with academic communities to build scientific understanding”
On October 26, 2026 (U.S. local time), OpenAI released a research announcement titled “Working with academic communities to build scientific understanding.” The announcement presented three new discoveries in harmonic analysis, computational complexity theory, and quantum field theory, spanning the fields of mathematics, computer science, and physics. I asked generative AI to investigate the implications of this announcement.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, do not necessarily reflect the actual state of affairs, and may contain errors. Please bear these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

OpenAI、「GPT-6.1 Terra」発表

28/10/2026

0 Comments

 
OpenAIは2026年10月26日(米国時間)、中位モデルの新モデル「GPT-6.1 Terra」を公開しました。画像や画面を見て判断する力、パソコンの画面を操作する正確さ、読みやすい文章を書く力が優れているモデルで、API料金は前の GPT-5.6 Terra の半分です。
「GPT-6.1 Terra」の知財業務への影響について生成AIに調べさせました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
OpenAI、コンピューター操作に強い「GPT-6.1 Terra」発表 入力100万トークン1ドル
https://www.sbbit.jp/article/cont1/187830
 
 
OpenAI Announces “GPT-6.1 Terra”
On October 26, 2026 (U.S. time), OpenAI released GPT-6.1 Terra, a new mid-tier model. The model excels at interpreting and reasoning about images and on-screen content, accurately interacting with computer interfaces, and producing clear, readable text. Its API pricing is half that of its predecessor, GPT-5.6 Terra.
I asked generative AI to investigate the implications of GPT-6.1 Terra for intellectual property (IP) work.
Please note that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect actual conditions, and may contain errors. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

「HP ZGX Fury G1n AI Station」の知財業務への影響

28/10/2026

0 Comments

 
日本HPは2026年10月26日、AI開発に特化したタワー型ワークステーション「HP ZGX Fury G1n AI Station」を発表しました。「NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip」を搭載し、大規模AIモデルの開発や推論処理をローカル環境で実行できます。機密性の高いデータを、ネットワークを閉ざした環境でAIを利用できる点が大きなメリットとなり、月に数十台規模の販売を見込んでいるようです。
「HP ZGX Fury G1n AI Station」の知財業務への影響について生成AIに調べさせました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
データセンター級のAIをローカルで、HPが大規模AI向けワークステーション発売
2026年10月26日
https://monoist.itmedia.co.jp/mn/articles/2610/26/news027.html
 
 
Implications of the HP ZGX Fury G1n AI Station for Intellectual Property Work
On October 26, 2026, HP Japan announced the HP ZGX Fury G1n AI Station, a tower workstation designed specifically for AI development. Equipped with the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, it enables users to develop large-scale AI models and run inference locally. A major advantage is the ability to use AI to process highly confidential data in a network-isolated environment. The company reportedly expects to sell several dozen units per month.
I asked generative AI to investigate the implications of the HP ZGX Fury G1n AI Station for intellectual property work.
Please note that the findings of this AI-generated research and analysis are based solely on publicly available information, do not necessarily reflect the actual situation, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

知財業務にAIエージェントを適用する際のハーネス設計

27/10/2026

0 Comments

 
大規模言語モデル(LLM)を知的財産業務(先行技術調査、明細書およびクレームのドラフティング、拒絶理由通知への応答、FTO調査等)へ適用する際、実務上の成否を分ける最大の要因はベースモデルのパラメータ規模ではなく、モデルを包摂するソフトウェア基盤である「エージェント・ハーネス(Agent Harness)」の設計にあると考えられます。
知財業務にAIエージェントを適用する際のハーネス設計についての注意点を生成AIに調べさせました。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
Harness Design for AI Agents in Intellectual Property Practice
When applying large language models (LLMs) to intellectual property (IP) work—including prior-art searches, drafting patent specifications and claims, responding to office actions, and freedom-to-operate (FTO) searches—the most critical factor determining success or failure in practice is likely to be the design of the “agent harness,” the software infrastructure that surrounds and supports the model, rather than the parameter count of the base model.
I asked generative AI to investigate the key considerations in designing harnesses for AI agents used in IP work.
Please note that the research and analysis conducted by generative AI are based solely on publicly available information, do not necessarily reflect actual circumstances, and may contain inaccuracies. Please bear these limitations in mind when reviewing the findings.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments
<<Previous

    著者

    萬秀憲

    アーカイブ

    April 2026
    March 2026
    February 2026
    January 2026
    December 2025
    November 2025
    October 2025
    September 2025
    August 2025
    July 2025
    June 2025
    May 2025
    April 2025
    March 2025
    February 2025
    January 2025
    December 2024
    November 2024
    October 2024
    September 2024
    August 2024
    July 2024
    June 2024
    May 2024
    April 2024
    March 2024
    February 2024
    January 2024
    December 2023
    November 2023
    October 2023
    September 2023
    August 2023
    July 2023
    June 2023
    May 2023
    April 2023
    March 2023
    February 2023
    January 2023
    December 2022
    November 2022
    October 2022
    September 2022
    August 2022
    July 2022
    June 2022
    May 2022
    April 2022
    March 2022
    February 2022
    January 2022
    December 2021
    November 2021
    October 2021
    September 2021
    August 2021
    July 2021
    June 2021
    May 2021
    April 2021
    March 2021
    February 2021
    January 2021
    December 2020
    November 2020
    October 2020
    September 2020
    August 2020
    July 2020
    June 2020

    カテゴリー

    All

    RSS Feed

Copyright © よろず知財戦略コンサルティング All Rights Reserved.
サイトはWeeblyにより提供され、お名前.comにより管理されています
  • Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法